Fig. 4

KNN workflow for a KNN classification application created using Lemonade. The data reader extract data from a file; specific features are extracted from records and then only the columns necessary for classification are selected. From that data, a sample is taken to feed the training model, which uses the KNN classifier as its engine. The trained model is then applied to the remainder of the data, and the Projection box provides a visualization of the result. The colors are used to identify modules that can be grouped during code generation (discussed in Section 5.3)